An offline backup is where the data or system that is backed up is unavailable for other uses for the duration of the backup. In database circles, these types of backups are referred to as cold backups.
As is implied by its name, this results in an outage on the data and application(s) using that data and will take as long as the backup takes to complete. This presents obvious disadvantages for 24/7 businesses that cannot be denied. However, in applications or businesses where 24 × 7 availability is not required, there are some advantages to offline backups, most typically where there are no dedicated IT staff to perform more complex recoveries that may arise from other forms of backup.
Such backups can be cheaper, too. For some enterprise products, database backups will be a licensed feature*: businesses may enact a policy whereby hot/online backups are performed for production databases, but development and test databases requiring backup are instead shutdown.
Potential Performance Problems
Offline backups may actually cause performance problems outside of the backup window. Many modern enterprise applications, particularly databases, use sophisticated caching techniques to reduce the number of IO operations required, with SAP and Oracle being two such applications. As soon as an application is shut down, any cache it may have been maintaining is typically lost. Indeed, some vendors will strongly recommend against frequent restarts of their applications for cache performance reasons.
Advantages of Offline Backups
The advantages of offline backups include the following:
- Recovery is typically trivial.
- For even complex environments, recoveries can be performed by staff who might otherwise not be application or system administrators when following well-tested instructions.
- When combined with snapshots, volume replication, and/or clustering, this technique might allow for simpler backups with minimal application downtime by performing the backup against a temporarily shut down and static copy or node.
The disadvantages of offline backups include the following:
- Without snapshots or other expensive techniques, applications that rely on data are unavailable for the duration of the backup.
- Care has to be taken to ensure that all components of the data are unused during backup. (For example, if data resides on multiple filesystems, there must be no changes to any of those filesystems during backup for the purposes of consistency.)
- For databases in particular, incremental offline backups are usually not supported—that is, an offline backup will force a new full backup. If the database is only a few gigabytes in size, this may be OK, however, it quickly becomes impractical as the size of the database increases.
- Over time, this model is unlikely to work as the amount of data to be backed up and the business systems availability requirements increase.
Published on Fri 14 October 2016 by Manny Larson in Computer Science with tag(s): backup